How to get normalized silhouette image from video

I want to get normalized silhouette image.

However, I got this images when I use the below code.

Figure 1 Figure 2

Figure 3 (I want to get this image that normalized at the center of the image)

function Y=normalize8(X)
%% Init. operations
X=double(X);
[a,b]=size(X);
%% Adjust the dynamic range to the 8-bit interval
max_v_x = max(max(X));
min_v_x = min(min(X));
Y=ceil(((X - min_v_x*ones(a,b))./(max_v_x*(ones(a,b))-min_v_x*(ones(a,b))))*255);
end

clear all
close all

RGB = imread('weighted17.jpg');


I = rgb2gray(RGB);
figure
imshow(I)

Y=normalize8(I)
figure
imshow(Y)

Answer 1

[Image Analyst]

Use

mask = bwareafilt(mask, 1);

to extract the largest blob.

Or take blobs just 64 pixels and bigger with this:

mask = bwareafilt(mask, [64, inf]);

But the BIG problem is that the image seems to be either an indexed or dithered image, or an image that had all the quality JPEG'ed out of it. NEVER use JPG images for image analysis. See if you can get the full resolution, uncompressed version of the image somehow. This image is virtually useless.

[Kong]

Thank you so much!

I got this error.

Also, I got the image from back substraction.

As you recommended, do I need to find other methods to extract images?

[Image Analyst]

Your BW is probably an RGB image. Since it's a JPG image, I'm almost certain it's a color image. Try taking just the red channel

BW = BW(:, :, 1) > 127; % Binarize the red channel

[Kong]

Thank you so much!

I did it successfully!

Can I move the human silhouette to the center of image?